whio-sdk

A JS SDK that interacts with our medical-assistant-api

Audio Transcription Queue

The SDK now supports automatic audio transcription through a background queue system. When you upload audio files, they are automatically processed and transcribed.

Upload Audio with Transcription Queue

import { ApiSDK, AudioFileStatus } from 'whio-sdk';

const sdk = new ApiSDK({ baseUrl: 'https://your-api.com' });

// Simple upload (automatically triggers transcription)
const audioFile = await sdk.uploadAudioFileToSession(sessionId, file);

// Upload with cultural context
const audioFileWithContext = await sdk.uploadAudioFileToSession(
  sessionId, 
  file, 
  'audio.mp3', 
  'This is cultural context for the audio'
);

// Alternative method with options object
const audioFileAlt = await sdk.uploadAudioFileWithTranscriptionQueue(sessionId, file, {
  fileName: 'recording.mp3',
  culturalTranscription: 'Patient discussing symptoms in Māori'
});

Check Transcription Results

// Get specific audio file with transcription results
const audioFile = await sdk.getAudioFile(audioFileId);

if (audioFile.status === AudioFileStatus.TRANSCRIBED) {
  console.log('Transcription:', audioFile.transcription);
  console.log('Cultural Context:', audioFile.cultural_transcription);
}

// Get all transcribed files
const transcribedFiles = await sdk.getTranscribedAudioFiles();

// Get files currently being processed
const processingFiles = await sdk.getProcessingAudioFiles();

// Get failed transcriptions
const failedFiles = await sdk.getFailedAudioFiles();

Audio File Status

Audio files have the following status values:

• UPLOADED - File uploaded, waiting for transcription
• PROCESSING - Currently being transcribed
• TRANSCRIBED - Transcription complete
• FAILED - Transcription failed

Key Features

• Automatic Processing: Upload triggers transcription automatically
• Serial Processing: Files are processed one at a time for reliability
• Cultural Context: Optional cultural transcription field for context
• Status Tracking: Monitor transcription progress
• Error Handling: Graceful handling of transcription failures
• Fire and Forget: No polling required - just upload and check later

Example Usage

// Upload and get immediate response
const audioFile = await sdk.uploadAudioFileToSession(sessionId, audioBlob, 'recording.wav');
console.log('Status:', audioFile.status); // 'UPLOADED'

// Check later for transcription results
setTimeout(async () => {
  const updatedFile = await sdk.getAudioFile(audioFile.id);
  if (updatedFile.status === AudioFileStatus.TRANSCRIBED) {
    console.log('Transcription ready:', updatedFile.transcription);
  }
}, 10000); // Check after 10 seconds